AI Trading Bots? Retail Beware

Blain’s Morning Porridge – May 31^th 2023: AI Trading Bots? Retail Beware

“Stocks are about guessing future outcomes, bonds are about understanding current returns.”

Let’s assume the US Debt crisis is averted. What comes next? A full-on speculative bubble and hype around Artificial Intelligence is underway. How will AI influence trading strategies, or turn markets into gray-goo? Will AI trading prove to be little more than finding new innovative ways to relieve retail investors of their money?

What next for markets? Let’s assume the US debt ceiling crisis has been now pushed down the road another couple of years – sure, lots of shouting still to come in Washington, but it’s done. What comes next in markets? More banking crisis (possible)? Another boom in stocks? Or rising rates trigger further corporate uncertainty? In the face of the many market uncertainties do we have a period of calm common sense or a further escalation of irrational exuberance?

I fear the latter – because that is how markets have learnt to behave. The next bubble is upon us and its called AI. Lord preserve us.. we’ve been here before…

I have a very simple outlook on the real value of equity. What are they really, really worth? When interest rates are low – as they were for the 13 years following the Global Financial Crisis in 2012 – the critical relationship between the prices of bonds and equities becomes distorted. If the risk-free rate set by government bonds is too low, investors seek to boost returns by looking elsewhere – ie stocks. Equity yields/returns became more attractive relative to bonds – explaining the real reason stocks rallied so strongly for a decade despite pretty lethargic growth and limited productivity gains across Western Economies.

Cheap money, and the sheer amount of it available (liquidity), meant investors flocked to stocks, triggering massive, (but largely uncommented upon), financial asset inflation through the QE era (2010-2022). It all got a bit silly.

Evidence? Please explain to me how stocks are not overvalued. From 2010 to 2022 US stocks rose 270%. Yet the GDP of the economy only rose 40%. How can stocks have increased in value at a rate 7 times faster than the value of the underlying economy? Seriously. Explain it to me… What a company is really, really worth is a factor of the profits it will create over time discounted by interest rates. As interest rates rise, that number falls. Doh. (Light bulb moment?)

I can explain why it happened in terms of the relative return distortions. Today, I am told by stock gurus that the US market is poised for further spectacular gains, despite the inflation risks, the interest rate rises and the possibility of global recession/geopolitical tensions and conflict. Markets look overvalued – but its only Tech mega-stocks going up. Everything else is flatline. That’s where fundamental value lies – in fair priced stocks.

Think about it… In 2006 Warren Buffet was regarded as the most successful investor alive. Decades later, and a host of failed celebrity stock-pickers later, Warren Buffet’s considered log-term value investing approach leaves him still the most successful and recognised investor around. All around him noise. He stuck to the basics of long-term fundamental value.

What we saw during the QE distortion period was a rise in speculation. Investors chased returns, they suffered FOMO (fear of missing out), and increasingly jumped into the narrative – whatever that was. They were consumed by the market noise – believing the hype. For the last 14 years it’s been a cataclysm of market froth and hype ranging from WeWork, Tesla, ARK, Crypto, SPACS, NFTs, Meme Stocks – the next new, new thing promising to make everyone spectacularly rich for minimal effort.

Look in your Linked-in Feed this morning and you will still find the nonsense being pumped into the investoverse – the 24 year old making $8mm a year day-trading with his 4 simple charts and investment rules beating the market every day, and he’s willing to tell you how. Financial shysters have been spitting cider into the ears of gullible stock shills since the concept of cash was invented….

The last 14 years of madness has programmed us to willingly believe in the next new, new thing.. We listen to anything that promises to make us rich for the least effort. No one seems to remember the last thing that promised to be the golden egg, but ended up a broken omelette. We fail to learn lessons from the past…. which is really surprising because we are supposedly living in an age of increasing knowledge and data that explains what went before.

The problem is we don’t actually take the time to understand. Believe first, understand later. The fastest way to lose money in markets is a lack of due diligence and believing what you don’t understand. Surely we can then avoid repeating the same mistakes? Nope. Market speculative froth and the curious behaviour of crowds is inescapable. AI is now on the froth list.

Yesterday I was asked a very simple question – how is Artificial Intelligence going to change trading?

Initially I thought this is a good one. Dangerous, but interesting. If everyone has an AI driven trading strategy that really CONNECTS and CORRELATES all the multiple data and behaviours that make up markets, then that’s going to kill trading as everyone will trade the same way.. after all it’s the fact I think stocks are overvalued and you think they are a bargain that creates a market. Not when everyone thinks the same way because the AI says so.

My grasp of tech has always been “tenuous” (I am not allowed to change plugs at home), but to build an AI Trading/Investment machine needs four things:

• Highly specialised and expensive Generative AI Chips to process and “understand” all the data and feeds.
• Loads of Dynamic Random Access Memory (great news for DRAM chip producers) to store and flow the data.
• Data – access to information, current feeds and historical data sets. (Remembering; no amount of information is enough to be “perfect”.)
• Being trained to understand what it all means.

Processing power means properly configured AI trade-bots will be able to better correlate and drive smarter hedging, futures and options, and structured trading returns. We will need fewer analysts data mining repetitive seams of information. The machines will read charts and spot inconsistency. They will be able to better and faster understand the maths behind how one complex options trade may trigger an arbitrage opportunity elsewhere. But even an AI as smart as a human mind will never be able to perfectly predict the future. Even the best AI trading strategy can’t foresee the no-see-ums, black swans, or unexpected crowd behaviours that drive markets.

And then I started digging. I looked on the net and found there are already a host of AI Stock Trading tools available – targeting the easy marks: retail investors.

I don’t believe for a moment there are true “artificial intelligences” already out there making trading decisions. There are powerful computers doing a lot of sums which “entreprenuers” shysters are passing off as smart AI trading platforms to lure gullible retail investors. The systems are not “intelligent”. At best they are machines running a load of current data versus historical data through algorithms looking for correlations, spotting anomalies and making a binary decision if it’s an up or down (buy or sell). They are just number crunchers – they are not smart, clever, or intelligent.

I make no judgements on the truth of the claims made by the AI’s referenced in this article: 10 “Best” AI Stock Trading Bots but they filled me with dread. Some are just trading tools, but others claim to be AIs, watching every tick, experienced development teams, scanning ETFs, leveraging off IBM Watson etc etc. One says it combines data from ETF and AI – which is as meaningless drivel as I’ve read in a long while. Another learns from users, but if the users are fools, then definitionally so is the system. Another is “AI software that enables you to build, backtest and automate trading strategies in Equities and Cryptocurrencies” yeah, sure it does.

In essence, these Investing AIs are unfiltered bollchocks. Are they really AIs that have sprung fully formed into the investment world? Or do they exist to help retail investors lighten their heavy wallets? You decide.

If they are real, then I bet some smart hedge fund manager is already training his AI to arbitrage them. All these scenarios raise a host of new issues for regulators to think about regulate – what is AI and how to stop AIs predating each other.

In the space of just a few months – since ChatGDP started filling the media streams – AI has become the most searched investment theme. It’s become a full-on circus. It’s now being spun as the ultimate trading dream – promising the golden key to effortless wealth creation, a perpetual motion investment machine that churns out market beating returns on a daily basis: “Buy my system, it will make you rich” is the basis on which the foolish are relieved of their cash.

Be aware. Be careful. Be suspicious.

Five Things To Read This Morning

FT                   Robert Armstrong: Citigroup: Cheap or Hopeless?

BBerg              Everywhere You Look In China Are Signs Of More Market Misery

BBerg              Cathie Wood Says Software Stocks Are Next AI Bet After Nvidia

WSJ                Where is the US Economy Headed? Follow the Money

Torygraph       Canary Wharf finances under pressure as workers shun the office

Out of time and back to the day job (but going for a swim in the river first…)

Bill Blain

Strategist – Shard Capital